This study proposes the use of Machine Learning models to predict the early onset of sepsis using deidentified clinical data from Montefiore Medical Center in Bronx, NY, USA. A supervised learning approach was adopted, wherein an XGBoost model was trained utilizing 80\% of the train dataset, encompassing 107 features (including the original and derived features). Subsequently, the model was evaluated on the remaining 20\% of the test data. The model was validated on prospective data that was entirely unseen during the training phase. To assess the model's performance at the individual patient level and timeliness of the prediction, a normalized utility score was employed, a widely recognized scoring methodology for sepsis detection, as outlined in the PhysioNet Sepsis Challenge paper. Metrics such as F1 Score, Sensitivity, Specificity, and Flag Rate were also devised. The model achieved a normalized utility score of 0.494 on test data and 0.378 on prospective data at threshold 0.3. The F1 scores were 80.8\% and 67.1\% respectively for the test data and the prospective data for the same threshold, highlighting its potential to be integrated into clinical decision-making processes effectively. These results bear testament to the model's robust predictive capabilities and its potential to substantially impact clinical decision-making processes.

A new sepsis screening tool developed using machine learning was timelier and more discriminating than several benchmark screening tools, according to data published in the Annals of Emergency Medicine. The new tool, the Risk of Sepsis (RoS) score, was developed using machine learning and compared with benchmark sepsis-screening tools such as the systemic inflammatory response syndrome, sequential organ failure assessment, quick sequential organ failure assessment, modified early warning score, and national early warning score. Investigators used retrospective electronic health record data from adult patients from 49 urban community hospital emergency departments over a 22-month period to derive and test the model. A total of 2,759,529 records were obtained using the Rhee, et al1 standard for clinical surveillance criteria as the definition of sepsis and the primary target for developing the model. The selection process consisted of 3 stages: (1) existing models for sepsis screening were reviewed, (2) consultation with local subject matter experts, and (3) supervised machine learning called gradient boosting.